Adversarial Domain Adaptation with Semantic Consistency for Cross-Domain Image Classification

In the cross-domain image classification scenario, domain adaption aims to address the challenge of transferring the knowledge obtained from the source domain to the target domain that is regarded as similar but different from the source domain. To get more reliable domain invariant representations, recent methods start to consider class-level distribution alignment across the source and target domains by adaptively assigning pseudo target labels. However, these approaches are vulnerable to the error accumulation and hence unable to preserve cross-domain category consistency. Because the accuracy of pseudo labels cannot be guaranteed explicitly. In this paper, we propose Adversarial Domain Adaptation with Semantic Consistency (ADASC) model to align the discriminative features across domains progressively and effectively, via exploiting the class-level relations between domains. Specifically, to simultaneously alleviate the negative influence of the false pseudo-target labels and get the discriminative domain invariant features, we introduce an Adaptive Centroid Alignment (ACA) strategy and a Class Discriminative Constraint (CDC) step to complement each other iteratively and alternatively in an end-to-end framework. Extensive experiments are conducted on several unsupervised domain adaptation datasets, and the results show that ADASC outperforms the state-of-the-art methods.

• Publication year

  2019

• Venue

  International Conference on Information and Knowledge Management

• Publication date

  2019-11-03

• Fields of study

  Computer Science

• Identifiers
  DOI 10.1145/3357384.3357918
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Generative Adversarial Networks

  2021cited by this paper
• M-ADDA: Unsupervised Domain Adaptation with Deep Metric Learning

  2018cited by this paper
• Multi-Adversarial Domain Adaptation

  2018influential reference
• Importance Weighted Adversarial Nets for Partial Domain Adaptation

  2018cited by this paper
• CoNet: Collaborative Cross Networks for Cross-Domain Recommendation

  2018cited by this paper
• Taking a Closer Look at Domain Shift: Category-Level Adversaries for Semantics Consistent Domain Adaptation

  2018cited by this paper
• Collaborative and Adversarial Network for Unsupervised Domain Adaptation

  2018influential reference
• A Deep Framework for Cross-Domain and Cross-System Recommendations

  2018cited by this paper
• Online Heterogeneous Transfer Metric Learning

  2018cited by this paper
• Content Recommendation by Noise Contrastive Transfer Learning of Feature Representation

  2017cited by this paper
• Adversarial Discriminative Domain Adaptation

  2017influential reference
• Asymmetric Tri-training for Unsupervised Domain Adaptation

  2017cited by this paper
• Transfer Adversarial Hashing for Hamming Space Retrieval

  2017cited by this paper
• Conditional Adversarial Domain Adaptation

  2017cited by this paper
• Label Efficient Learning of Transferable Representations acrosss Domains and Tasks

  2017cited by this paper
• Prototypical Networks for Few-shot Learning

  2017cited by this paper
• Few-Shot Adversarial Domain Adaptation

  2017cited by this paper
• Central Moment Discrepancy (CMD) for Domain-Invariant Representation Learning

  2017influential reference
• Adversarial Feature Augmentation for Unsupervised Domain Adaptation

  2017cited by this paper
• Deep Transfer Learning with Joint Adaptation Networks

  2016influential reference
• Sentiment Domain Adaptation with Multi-Level Contextual Sentiment Knowledge

  2016cited by this paper
• Deep Reconstruction-Classification Networks for Unsupervised Domain Adaptation

  2016cited by this paper
• Deep CORAL: Correlation Alignment for Deep Domain Adaptation

  2016cited by this paper
• Deep Residual Learning for Image Recognition

  2015influential reference
• Learning Transferable Features with Deep Adaptation Networks

  2015influential reference
• Simultaneous Deep Transfer Across Domains and Tasks

  2015cited by this paper
• Domain-Adversarial Training of Neural Networks

  2015influential reference
• An Informative Logistic Regression for Cross-Domain Image Classification

  2015cited by this paper
• Deep Domain Confusion: Maximizing for Domain Invariance

  2014cited by this paper
• Unsupervised Domain Adaptation by Backpropagation

  2014influential reference
• Reading Digits in Natural Images with Unsupervised Feature Learning

  2011influential reference
• Adapting Visual Category Models to New Domains

  2010cited by this paper
• A theory of learning from different domains

  2010influential reference
• A Kernel Method for the Two-Sample-Problem

  2006influential reference
• Gradient-based learning applied to document recognition

  1998influential reference
• A Database for Handwritten Text Recognition Research

  1994cited by this paper

• FALCON: Few-Shot Adversarial Learning for Cross-Domain Medical Image Segmentation

  2026cites this paper
• Pseudo-labeling Integrating Centers and Samples with Consistent Selection Mechanism for Unsupervised Domain Adaptation

  2023cites this paper
• Synthetic-to-Real Domain Adaptation for Action Recognition: A Dataset and Baseline Performances

  2023cites this paper
• Discriminative transfer feature learning based on robust-centers

  2022cites this paper
• Domain adversarial tangent subspace alignment for explainable domain adaptation

  2022cites this paper
• Adapted human pose: monocular 3D human pose estimation with zero real 3D pose data

  2021cites this paper
• Norma: A Hybrid Feature Alignment for Class-Aware Unsupervised Domain Adaptation

  2021cites this paper
• Knowledge Preserving and Distribution Alignment for Heterogeneous Domain Adaptation

  2021cites this paper
• Soft Labels Transfer with Discriminative Representations Learning for Unsupervised Domain Adaptation

  2020cites this paper